We add new devices to NPM and NCM using the API from our CMDB, so they can be up/down monitored.
- the LLDP poller is added to nodes so LLDP data is collected by NPM
We use NCM to periodically perform a Node Inventory, this collects all the interfaces on a node.
We then have a report that combines these two, finds matching physical interfaces, and subtracts any monitored interfaces to highlight interfaces that need monitoring turned up:
Here is the SQL query:
SELECT distinct n1.ip_address, i4.interfaceindex, i4.interfacename,RemoteSystemName, n1.nodeid,n2.nodeid as rnodeid
FROM NodeLldpEntries NLE inner join nodes N1
on NLE.nodeid=n1.nodeid
inner join nodes N2
on NLE.remotesystemname=n2.sysname
inner join NCM_NodeProperties n3
on n3.corenodeid=n1.nodeid
inner join NCM_Interfaces i3
on i3.interfaceindex = nle.localportnumber and i3.nodeid=n3.nodeid
inner join NCM_Interfaces i4
on i3.macaddress = i4.macaddress and i4.PhysicalInterface='Y'
left outer join Interfaces i
on i.interfaceindex = nle.localportnumber and i.nodeid=nle.nodeid
left outer join Interfaces i5
on i5.interfaceindex = i4.interfaceindex and i5.nodeid=nle.nodeid
where i.interfaceid is null and i5.interfaceindex is null
and n1.caption not like '%@%' and n1.unmanaged=0 and n2.unmanaged=0
and i4.interfacename = LEFT(i3.interfacename,DATALENGTH(i3.interfacename)-2)
order by 1,2
This is significantly easier than running discovery and allows for rules-based monitoring decisions.
The next step will be to migrate the query to SWQL, code this into a script that runs each day, and automatically turn up monitoring according to defined rules.
(I'm hiring a Network Engineer to also look after our Orion install; if this kind of thing fascinates you, and you want to work in the PNW, check out jobs at uw.edu)